How Inflationary are Oil Price Shocks? A Regional Analysis Jon Christensson*

advertisement
Proceedings of the 5th Annual GRASP Symposium, Wichita State University, 2009
How Inflationary are Oil Price Shocks? A Regional Analysis
Jon Christensson*
Department of Economics. W. Frank Barton School of Business
(1).
Abstract. The impact of oil shocks is analyzed by estimating
an augmented Phillips curve on a national, regional and city
level in the United States. A significant pass-through to
inflation (including all items) is recorded for all regions, while
core inflation remains largely muted. The West region has
experienced a much lower pass-through than other regions
and a few reasons for this are; greater oil efficiency, lower
inflation variability and a lower exchange rate pass-through in
the West. Also noted is an increasing trend for pass-through
to inflation since the late 1980’s, and the contrary was found
for core inflation.
(1)
is the seasonally adjusted inflation rate,
is the unemployment gap created by
subtracting the NAIRU (Non-Accelerating Inflation
Rate of Unemployment) from the actual seasonally
adjusted unemployment rate,
is the UK Brent oil
price in U.S. dollars. The inflation rate and the
unemployment data was collected from BLS while the
1. Introduction
Intuitively, when the oil price increases due to a
negative supply shock, the input cost of firms increases.
If workers are rational, they will adjust their
inflationary expectation and demand higher wages
leading to higher labor costs. This causes the aggregate
supply curve to shift to the left and prices to increase.
Hence, the oil price has implicitly passed on its price
hike to the average price level in the economy.
oil price comes from IFS. The
polynomial in the lag operator.
implies a
The paper will use four models. Model 1 will have a
fixed lag structure and polynomial degree of 6 and 2,
respectively. Model 2 will have a fixed lag structure and
polynomial degree of 12 and 2, respectively. Model 3
will have a varying lag structure and polynomial
degree. Model 4 computes a time-varying pass-through
coefficient allowing for the varying lag structure and
polynomial degree of model 3.
There have been surprisingly few papers specifically
investigating the oil price pass-through to inflation. The
existing studies have largely noted a decline in the passthrough in recent times. A few reasons identified for
this decline include globalization, more flexible labor
markets, reduction in oil intensity, and a declining
exchange rate pass-through. [1] [2] [3]
The reason for inclusion of the first two models is to be
able to compare the pass-through across regions as well
as across price categories. Two different lag structures
are imposed as to not over/under fit the model which is
a serious issue. This is the primary reason of the third
model; by determining the optimal lag and polynomial
structure, the issue of over/under fitting the model is
side-stepped. The fourth model is put forth to analyze
how the pass-through has changed over time since the
late 1980’s.
The main contribution of this paper is to investigate the
pass-through on regional and disaggregated levels, and
how it changes over time. More precisely, this paper
measures the oil price pass-through to various
consumer price categories on a national, regional and
city level in the United States.
Incorporating the various models above separates this
paper from the rest. Hooker [5] uses the same lag
structures as Fuhrer [4] which allows for different lags,
though they do not specify to what degree of
polynomial. De Gregorio, Landerretche and Neilson [3]
claim to follow Fuhrer [4] and Hooker [5] but they do
not mention the use of a polynomial in the lag operator.
2. Experiment, Results, Discussion, and Significance
This paper follows the framework of Fuhrer [4],
Hooker [5] and De Gregorio, Landerretche and Neilson
[3], though with some differences which are discussed
shortly. An estimation of an augmented Phillips curve
is estimated and the general form is shown in equation
22
Proceedings of the 5th Annual GRASP Symposium, Wichita State University, 2009
It seems as if they impose a regular lag structure and fix
this lag to four quarters due to comparability issues.
Model 4 reveals an increasing trend of pass-through to
inflation (including all items) on a national and regional
level, while a negative trend is recorded for the core
inflation.
For most series the time period stretches from January
1987 to September 2008. Lastly, De Gregorio,
Landerretche and Neilson [3] use the output gap, while
Fuhrer [4] and Hooker [5] use the plain unemployment
rate as a measure of economic activity. Due to the
regional analysis in this paper the unemployment gap
must be used as a proxy for economic activity.
The most striking finding is the lower pass-through in
the West. One possible explanation to the lower passthrough is a decline in oil intensity. The West uses less
oil to produce the same amount of GDP as other
regions. A second explanation is the fact that the West
has a smaller exchange rate pass-through and thirdly,
that the standard deviations of many price series are
lower in the West compared to other regions.
From equation (1) the pass-through coefficient is
derived and shown below in equation (2).
∑
∑
For future research it may be beneficial to investigate
how the various regions’ economic structure impacts
the oil price pass-through.
(2)
where the summation goes from i to L (in this case will
be either 6 in model (1), 12 in model (2) and varying in
model (3) and (4)).
3. Conclusions
This paper differentiates itself from recent literature by
focusing on the oil price pass-through on a regional
level, as well as on a disaggregated level. A general
Phillips curve is estimated on a national, regional, and
city level. A significant pass-through is recorded in
regular inflation and other price categories that directly
incorporate oil, while the core inflation remains muted.
The most striking result can be seen in the West where
almost all significant price categories have a lower
pass-through compared to the other regions. A few
explanations to this are higher oil use efficiency, lower
inflation variation and a lower exchange rate passthrough in the West. A rolling pass-through coefficient
was also computed where the oil price pass-through to
inflation has had an increasing trend since the late
1980’s, contradicting previous research. However, the
core inflation has seen a negative trend which is in line
with other research.
On a national level, previous studies found that after the
1980, the pass-through to the general price inflation
was around three percent, which understates the
findings of this paper where the estimated pass-through
in model 1 and 2 is 2.6% and 5.7% respectively [3]. In
other words, if the oil price increases by 1%, the
inflation is expected to increases by 0.03% (0.0260.057% according to the current findings). Hooker [5]
claims that a doubling of the oil price leads to
approximately a 1% direct increase in inflation.
Interestingly, the core consumer prices seems to have
been unaffected by changes in the oil price in both lag
specifications. Not surprisingly, one sees the largest
pass-through rates in energy, transportation, commodity
and nondurables prices that in one way or another
include either gasoline or some form of oil directly.
At a regional level, the Northeast has the highest passthrough to consumer prices including all items and also,
on average, the most price categories with higher passthrough. The most striking result is the Western region
which experiences the smallest pass-through in almost
every price category. Hence, an oil price shock would,
on average, influence the inflation in the West less than
it would have in other parts of the US.
4. Acknowledgements
Thanks to Dr. Jen-Chi Cheng for advice, as well as
Marcus Christensson for programming assistance.
[1] Rogoff, K. (2003). “Globalization and global disinflation”,
Economic Review, Federal Reserve Bank of Kansas City,
Proceedings, pp. 77-112
[2] Blanchard, O., and Galí, J. (2007). “The Macroeconomic Effects
of Oil Price Shocks: Why are the 2000s so different from the
1970s?”, MIT Department of Economics, Working Paper, No 0721
[3] De Gregorio, J., Landerretche, O., and Neilson, C. (2007).
“Another Pass-Through Bites the Dust? Oil Prices and Inflation”,
Working Paper, Central Bank of Chile, No. 417
[4] Fuhrer, J. (1995). “The Phillips Curve is Alive and Well” New
England Economic Review, March/April, pp. 41-56
[5] Hooker, M. (2002). “Are Oil Shocks Inflationary? Asymmetric
and Nonlinear Specification versus Changes in Regime”, Journal
of Money, Credit and Banking, 34(2), pp. 540-561
At a city level there is no clear consensus as of where
the highest pass-through is. However, yet again the
West region has, on average, the lowest pass-through
and least significant categories.
Model 3 gives comparable results and the main
important observation is the lower pass-through in the
West.
23
Download